Proceedings: 2003 DEER Conference
U.S. Department of Energy 9th Diesel Engine Emissions Reduction Conference
Newport, Rhode Island, August 24 to 28, 2003
SELECTIVE CATALYTIC REDUCTION OF DIESEL ENGINE NOX EMISSIONS USING ETHANOL
AS A REDUCTANT
Michael D. Kass, John F. Thomas, Samuel A. Lewis, Sr.,
John M. Storey, Norberto Domingo, and Ron L. Graves
Oak Ridge National Laboratory
ABSTRACT
NOx emissions from a heavy-duty diesel engine were
reduced by more than 90% and 80% utilizing a full-scale
ethanol-SCR system for space velocities of 21000/h and
57000/h respectively. These results were achieved for catalyst
temperatures between 360 and 400oC and for C1:NOx ratios of
4-6. The SCR process appears to rapidly convert ethanol to
acetaldehyde, which subsequently slipped past the catalyst at
appreciable levels at a space velocity of 57000/h. Ammonia
and N2O were produced during conversion; the concentrations
of each were higher for the low space velocity condition.
However, the concentration of N2O did not exceed 10 ppm. In
contrast to other catalyst technologies, NOx reduction appeared
to be enhanced by initial catalyst aging, with the presumed
mechanism being sulfate accumulation within the catalyst. A
concept for utilizing ethanol (distilled from an E-diesel fuel) as
the SCR reductant was demonstrated.
INTRODUCTION
The use of hydrocarbons to reduce diesel exhaust NOx
emissions via selective catalytic reduction (SCR) has not been
studied as extensively as urea-based SCR systems. Benchscale evaluations of hydrocarbon-SCR (HC-SCR) have typically
not demonstrated high efficiencies (greater than 80 percent NOx
reduction) for realistic space velocities (1). On the other hand
urea-SCR was developed in the 1970s for stationary
applications (including stationary diesel engines) and has
consistently demonstrated NOx reduction efficiencies greater
than 90 percent when the SCR system operates in a favorable
temperature window (2). In fact, urea-SCR is one of a few
catalyst technologies capable of reducing diesel NOx emissions
to the 2007 NOx limit of 0.2 g/bhp-hr (2).
There are also concerns with applying urea-based SCR to
on-road vehicles, including 1) the need for a separate onboard
tank for urea, 2) the infrastructure to supply urea, 3) residue
buildup during inadvertent over-injection or injection at low
temperatures, 4) corrosivity associated with urea, 5) high freeze
point, and 9) the performance is best at optimum NO/NO2 ratios
which, in turn, require a pre-oxidation catalyst to maintain
optimum NO/NO2 ratios which can only be partially controlled
using a pre-oxidation catalyst. Because of these concerns,
there is a need for more convenient SCR systems, preferably
those that use diesel fuel or a fuel-borne additive as the
Alexander Panov
Caterpillar, Inc.
reducing agent. As a result, investigators have been pursuing
new methods of HC-SCR.
HYDROCARBON SCR
Although the development of hydrocarbon-based SCR has
not progressed as far as urea-based systems, catalysts utilizing
hydrocarbon reductants have been developed and evaluated
since the 1980s. The NOx reduction efficiencies for HC-SCR
have tended to be lower than for ammonia-based SCR systems
(1, 3-7).
Of the HC-SCR systems studied, the alumina supported
(highly-loaded) silver catalysts have been identified as one of
the more promising NOx control technologies for light-duty
diesel emissions (1). Successful reducing agents include
paraffins, alcohols, and aldehydes.
In some cases, the
conversion levels can be greater than 80% for the temperature
range of 350 to 500oC and these systems have demonstrated
good resistance to both water and SO2 inhibition. Although the
effects of SO2 are not well known, recent investigations have
suggested that the formation of sulfate on the silver surfaces
actually improves the NOx conversion of these catalysts (8-9). It
is believed that the accumulation of sulfate on the silver sites
increases the selectivity toward NOx reduction, which
correspondingly increases HC slip and reduces N2O formation
(9).
ETHANOL AS A REDUCTANT
Ethanol has been demonstrated to be a successful
reducing agent for silver-doped alumina catalyst systems (1, 5,
8, 9). In the mid 1980s Caterpillar developed and marketed an
ethanol-SCR system for use with stationary diesel engines; NOx
conversion efficiencies were around 75% (10). More recent
studies have shown that when using ethanol as a reductant over
alumina supported silver catalysts, NOx conversion efficiencies
greater than 90% were observed for a C1:NOx ratio of 4 and a
space velocity (SV) of 50000/h (10). In addition it appears that
the NOx conversion efficiency is independent of the NO/NO2
ratio (8-9). These investigations; however, were performed
under controlled laboratory settings and actual performance
data from engine tests are lacking.
Ethanol is also of interest because it can be easily
incorporated into diesel fuel using an emulsifying agent.
Ethanol has been evaluated as an additive to diesel fuel over
1
the past 20 years to 1) lessen the demand for imported oil, 2)
promote the use of a domestic renewable resource to power
diesel engines, and 3) to lower particulate matter (PM)
emissions. Previous efforts have shown that ethanol-diesel
mixtures (commonly known as E-diesel) containing up to 15%
ethanol are relatively stable and do not adversely affect engine
operation (12).
The ethanol, ECD-1 fuel, and GE Betz additive (Blending
Agent DMX10011) were splash-blended by Growmark Inc. at
their manufacturing facility to form a stable micro-emulsion
containing 15 % by volume ethanol.
REDUCTANTS
Fuel-borne reductant separation
RATIONALE OF CURRENT WORK
The primary rationale guiding this effort was to investigate
the NOx reduction potential for ethanol-SCR using a heavy-duty
diesel engine. This included performing a convincing proof-ofconcept for use of a fuel-borne reductant. Specifically, we
wanted to show that the ethanol (contained within E-diesel) has
the potential to reduce NOx emissions via HC-SCR after being
separated (by mild distillation) from the fuel blend and used as
an injected reductant preceding an alumina supported silver
catalyst. An on-board processing system (not part of this work)
can be envisioned to strip ethanol from the base fuel and inject
it into the exhaust as a reductant for NOx conversion.
.
NOMENCLATURE
C1
DOE
ECM
E-diesel
EGR
FG
FTIR
GC-MS
HC
HC-SCR
NOx
NTRC
ORNL
PM
SCR
SV
A means for expressing hydrocarbon
concentration. It is obtained by multiplying
the number of carbon atoms per molecule by
the concentration in ppm.
Department of Energy
Electronic Control Module
A ethanol and diesel fuel mixture, usually
containing a blending agent and mixed as a
microemulsion
Exhaust Gas Recirculation
Fuel Grade
Fourier Transform Infrared
Gas Chromatigraph-Mass Spectrometer
Hydrocarbon
Hydrocarbon – Selective Catalytic Reduction
Oxides of Nitrogen
National Transportation Research Center
Oak Ridge National Laboratory
Particulate Matter
Selective Catalytic Reduction
Space Velocity
MATERIALS AND EQUIPMENT
TEST FUELS
The test fuels investigated in this study included BP/ARCO
ECD-1 and an E-diesel blend originally containing 15 vol.%
ethanol, 1.5 vol.% blending agent, and 83.5 vol.% ECD-1.
ECD-1 is a low sulfur diesel fuel (less than 15 ppmv) developed
to reduce the levels of regulated emissions. The ethanol used
in this study was fuel-grade (supplied by Williams-Pekin, Inc.)
and is a corn-derived product denatured using gasoline and
containing a corrosion inhibitor; pertinent specifications are
listed in Table 1.
Table 1. Specifications for fuel-grade ethanol supplied by
Williams-Pekin, Inc.
Ethanol content, vol.% 92.1 min
Methanol content, vol.%
0.5 max
Denaturant content, vol.%
2 min, 5 max
Water content, mass%
~0.5
Approximately 1 liter of ethanol was stripped from 7 liters of
the E-diesel blend through the application of moderate heat and
vacuum using a laboratory distillation device (Buchi RE-121
Rotovapor). Visual inspection showed that the distillate was of
single phase. This suggests that the distillate was primarily
ethanol since diesel fuel, including the light aromatic
constituents, is immiscible with ethanol, even at low
concentrations. This relatively simple test demonstrates the
feasibility of using an on-vehicle device to remove lesser
portions of ethanol from E-diesel.
The remaining ~6 liters of stripped E-diesel fuel (mostly
depleted of ethanol) was added back to the original blend which
subsequently lowered the ethanol content to ~13 vol.%. This
partially stripped E-diesel was also evaluated as a test fuel.
Evaluated reductants
The two reductants evaluated in this study were fuel-grade
ethanol in the as-received condition (untreated) and the ethanol
portion that had been stripped (distilled) from the E-diesel blend.
Both the untreated and stripped ethanol batches were evaluated
as reducing agents while running the engine on ECD-1.
Following these tests, the engine was operated using the 13
vol.% E-diesel blend while using the stripped ethanol as the
reductant. A single test was performed using reagent-grade
(high purity) ethanol.
SCR SYSTEM IMPLEMENTATION
The ethanol delivery system consisted of a cooled Buick
gasoline injector, a fuel injection pump (Mallory Series 60FI),
and a return style regulator, which was used to maintain
pressure near 400 kPa. The injector was located in a bend in
the exhaust 64 cm from the turbo outlet.
Caterpillar, Inc. provided two seven liter catalysts to ORNL. The
catalysts had a cell density of 31 cells/cm2 (200 cells/in2) and
measured 24.1cm (9.5 in) in diameter by 15.2 cm (6 in) long.
They were installed in the exhaust at a distance of ~1.5 meters
from the engine turbo outlet. During the initial test runs, the
second brick was installed directly behind the first catalyst.
During this investigation the exhaust was unfiltered (no
particulate traps were used).
EXPERIMENTAL FACILITY
The evaluation was conducted at the National Transportation
Research Center (NTRC) by ORNL staff. A Cummins 5.9 liter
ISB turbo-charged, direct injection, diesel engine (1999 model,
24 valve, in-line 6 cylinder) was used as the test engine. This
engine was altered from the commercially available version due
to the addition of a Cummins supplied exhaust gas recirculation
system and a non-standard ECM, fuel pump, fuel injectors, and
turbine. The engine was rated for 175 HP and 2.5 g/HP-h of
NOx over the heavy-duty STP. The EGR valve could be
controlled either via the ECM map or by manual override. The
engine was coupled to a General Electric direct current motoring
2
dynamometer capable of absorbing 224 kW (300 hp) and
motoring 213 kW (285 hp).
The ethanol-SCR system layout and sample locations are
shown schematically in Fig. 1. Gaseous emissions were
sampled from the raw exhaust stream and directed to a
standard emissions bench and a Fourier Transform Infrared
(FTIR) spectrometer. The standard bench (composed of Horiba
Ltd. and California Analytical Instruments analyzers) provided
measurements of NOx, THC, CO, CO2, and O2. The FTIR
spectrometer (Nicolet Instrument Corp. Magna-IR 560) provided
speciation of ammonia, N2O, and acetaldehyde.
using fuel-grade ethanol by injecting into a graduated cylinder
for measured time periods.
EXPERIMENTS
INITIAL SHAKEDOWN AND CATALYST DEGREENING
Both catalysts were initially mounted in the exhaust. Initial
shakedown evaluations were performed to evaluate the
operation and performance of the ethanol delivery system,
bench analyzers, and data acquisition system. The engine was
run at several different operating modes for about 2 to 3 hours
with reductant injection to briefly examine the performance of
the catalysts in the green state. Following this, the catalysts
were aged for approximately 10 hours while maintaining the
engine exhaust temperature near 400oC.
REDUCTANT INJECTION
The general method for testing at a given condition (set engine
speed and load, fuel type, reductant type, catalyst) was to begin
with a relatively low reductant injection rate. After achieving
steady-state conditions and recording data, the injection rate
was increased. In this manner we would sweep across a
reasonable range stoichiometry and stop when relatively high
HC slip was observed.
SPACE VELOCITY COMPARISON
Fig. 1. Schematic diagram showing layout of ethanol-SCR
components and sampling locations.
A stream of exhaust was sent to a micro-dilution tunnel for
additional speciation via GC-MS and for measurement of N2O
and ethanol slip. Bag samples were drawn from the microdiluter for hydrocarbon speciation via GC-MS. In addition a
photoacoustic spectrometer (from Innova AirTech Instruments
model 1312) was used to measure both ethanol slip and tailpipe
N2O emissions. This instrument has been used successfully to
quantify ethanol emissions from vehicles operating on
ethanol/gasoline blends (6).
Fuel consumption was measured using a Max Machinery Inc.
fuel flow measurement system. Intake airflow was measured
using a Marion laminar flow element. The space velocity of the
gas flux was calculated and reported for standard conditions.
A custom PC-based system was used to control reductant
injection. Injector duty cycle (% of time in open position) was
adjusted to control the flow rate of ethanol into the exhaust.
Other system parameters were held constant, such as the
injector pulsing frequency (10 Hz) and the ethanol injection line
pressure (~400 kPa) and ethanol temperature (30-40oC). The
injection system was calibrated for a relevant duty cycle range
During the initial test results after aging we suspected that the
first catalyst brick was providing the bulk of the SCR capability
and that the second catalyst brick was relatively ineffective. For
this reason, and also to achieve SV values in a realistic range,
the downstream (or second) catalyst brick was removed. For
most tests the ethanol-SCR system performance was evaluated
for a relatively low SV of ~21000/h and a higher SV near
57000/h, while attempting to maintain the same narrow catalyst
temperature range for both conditions. The engine operating
modes corresponding to these two conditions are detailed below
in Table 2. (SV is reported for STP.)
Table 2. Key engine operating parameters for space velocity
evaluation
Space
velocity
(1/h)
~21000
~57000
Engine
Speed
(rpm)
1115
2225
Engine
torque
(ft-lbs)
230
180
Catalyst
temperature
(oC)
350 to 380
350 to 380
NOx
flux
(g/min)
1.5
2.1
During normal operation the NOx flux for the low-speed
condition was significantly higher than the high-speed condition.
The EGR valve was controlled to be partially open for the low
speed (21000/h) condition in order to reduce the NOx emissions.
(During both operating modes the EGR valve is normally
closed.) As revealed in the table, the NOx flux ended up being
higher for the high-speed condition.
TEMPERATURE EVALUATION
In addition to measuring the conversion efficiency of the catalyst
between 350 to 400oC, data was also taken to evaluate
o
performance between 250 and 350 C. The torque applied to
3
the low-speed condition was reduced to obtain lower catalyst
core temperatures. The conversion efficiency was evaluated for
two C1(2X ethanol):NOx ratios.
EVALUATION OF REDUCTANT AND FUEL TYPE
The performance of the SCR catalysts was determined as a
function of the reductant type (untreated or distilled ethanol) for
both the low and high-speed conditions while operating the
engine using ECD-1 fuel. In addition the performance was also
evaluated (for both conditions) using distilled ethanol as the
reductant while operating the engine using the E-diesel blend
(which was the source of the distilled ethanol). In order to gain
insight into reductant impurity contributions, an additional
experiment was performed utilizing high-purity reagent-grade
ethanol with the engine operating on ECD-1 fuel. The NOx flux
and catalyst temperature did not change appreciably when
running the engine on ECD-1 or E-diesel at the above two
conditions. Bag samples for the GC-MS analysis were taken for
several of the test sequences. Generally bag samples were
taken for a perceived high and low slip condition (low and high
reductant injection rates).
installation are also included in the diagram, since this data
represents the initial state of the catalysts following the degreening period. (Note that because two catalyst bricks were
used for this condition, the SV is one-half the value of the other
presented data.) During the runs taken on October 8, 9, and 17
untreated fuel-grade (FG) ethanol was used as the reductant
and the engine was operated using ECD-1 fuel. The engine
was run using ECD-1 for the days of October 10 and 15, but
distilled FG ethanol was used as the reducing agent for these
evaluations. The Oct. 16 data was taken while running the
engine on E-diesel and utilizing distilled FG ethanol as the
reductant. This data set represents a simulation of a fuel-borne
reductant, where the fuel used to run the engine was also used
to supply the reducing agent for the catalyst. An additional data
set taken on Oct. 17 using high purity reagent-grade ethanol is
shown in Fig. 2.
Table 3.
condition
Test
RESULTS AND DISCUSSION
DESCRIPTION OF TEST CONDITIONS
The test conditions depicted in Table 3 show the key operating
parameters associated with each test condition, including the
day each particular study was performed.
Operating parameters associated for each test
Space
Velocity
date
shakedown tests
Fuel
ECD-1
10 hrs, catalyst aging
ECD-1
1
Oct. 7, 8
10,500
ECD-1
2
Oct. 8
28,500
ECD-1
3
21,000
ECD-1
57,000
ECD-1
21,000
ECD-1
As shown in Table 3, we evaluated 8 conditions as part of this
investigation. The test conditions labeled as 1, 3, 4, and 5 were
repeated on later days.
Key parameters included space
velocity, fuel type, and reductant type.
4
5
Oct. 8-10,
17
Oct. 8, 15,
17
Oct. 10, 15
ENGINE OPERATION AND EMISSIONS
6
Oct. 16
21,000
E-diesel
7
Oct. 16
57,000
E-diesel
8
Oct. 17
21,000
ECD-1
Analysis of the engine-out emissions for both engine operating
conditions is shown in Table 4. Both the chemiluminescent
bench detectors and the FTIR spectrometer showed that the
NOx in the exhaust was primarily NO.
The engine-out NOx emission levels were similar for both fuel
types. However, the E-diesel fuel did produce a small increase
in the emissions of HC and CO, which was consistent with
earlier results when running the engine on this fuel type. There
was also a drop in engine performance associated with E-diesel
usage. This is expected since the energy density of the 13%
ethanol fuel is about 6% lower than the straight ECD-1 fuel. In
addition, the combustion characteristics for the two fuels are not
identical.
NOx CONVERSION RESULTS
The NOx conversion efficiencies associated with the ethanolSCR system were measured over a 2-week period in October
2002.
The initial runs incorporated both catalyst bricks.
However the back brick was removed since it lowered the SV to
unrealistically low levels and made interpretation of the data
more difficult. As shown in Figs 2 and 3 the NOx conversion
efficiency increased with increasing C1:NOx ratio for both the
low SV condition (Fig. 2) and the high SV condition (Fig. 3).
The conversion efficiencies associated with the double brick
EtOH Reductant
Fuel-grade, as
received
none
Fuel-grade, as
received
Fuel-grade, as
received
Fuel-grade, as
received
Fuel-grade, as
received
Distilled from Ediesel
Distilled from Ediesel
Distilled from Ediesel
Reagent grade
Table 4. Bench analyzer results for engine operating conditions
and fuel type
Bench
Results
1115 rpm/230 ft-lbs
(21000/h)
ECD-1
E-Diesel
2225 rpm/180 ft-lbs
(57000/h)
ECD-1
E-Diesel
NOx, ppm
280-315
306-309
150-164
159-164
CO2, %
10.1
10.1
7.4-7.7
7.1-7.3
HC, ppm
12-23
24-29
18-27
39-46
CO, ppm
253-320
213-216
107-135
170-172
Therm. Eff.,
%
BMEP, psig
38.1
36.0
31.6
29.7
95.5-97.5
97.4
73-76
75.5
The data presented in Figs. 2 and 3 show that NOx conversion
increased with increasing ethanol delivery (C1:NOx ratio) for all
4
Because the catalyst performance improved with the sequence
of experimentation, we had difficulty making detailed
assessments of SCR performance with fuel type or reductant
grade. However, the Oct. 17 data presented in Fig. 2 shows no
observable difference in NOx conversion performance when
using either fuel-grade (untreated) or reagent-grade ethanol as
the reducing species. Likewise, the Oct. 16 and 17 results
shown in Fig. 3 indicate that NOx conversion was not greatly
affected when running the engine on E-diesel and using distilled
ethanol as the reducing species.
The Oct. 16 results clearly demonstrate the utility of using
ethanol as a fuel-borne reductant to reduce NOx emissions.
The NOx conversion was over 90% for a C1:NOx ratio of ~5 at
21000/h and near 85% for a C1:NOx ratio of 7 at 57000/h.
100
90
NOx Conversion (%)
80
70
FUEL PENALTY
For the experiments represented in Figs. 3 and 4, the C1:NOx
ratio can also be expressed as a fuel penalty. Figure 4 gives
calculated fuel penalty on an energy basis as a function of the
C1:NOx ratio for the two engine conditions (Table 2.). It was
assumed the ethanol reductant that is distilled from the fuel
contains 63% of the energy on a weight basis compared to
ECD1 fuel.
100
90
NOx Conversion (%)
80
70
60
50
40
Oct. 8: ECD1 fuel, Reductant = FG ethanol (untreated)*
30
Oct. 8: ECD1 fuel, Reductant = FG ethanol (untreated)
20
Oct. 15: ECD1 fuel, Reductant = FG ethanol (untreated)
10
Oct. 16: E-diesel fuel, Reductant = FG ethanol (distilled)
Oct. 17: ECD1 fuel, Reductant = FG ethanol (untreated)
0
0
2
4
6
8
10
12
14
16
C1(2X ethanol):NOx Ratio
Fig 3. NOx conversion as a function of C1:NOx ratio for the high
speed engine operating condition (2225 RPM, 182 ft-lb,
~57000/h space velocity. *Note, for this condition SV 26500/h.
3.5
3.0
Fuel Penalty, Energy Basis (%)
operating conditions. These figures also show that the level of
NOx conversion increased with the progression of
experimentation up to Oct 16 for both operating conditions. This
effect did not appear to be influenced by either fuel or reductant
type. For the low-speed condition (Fig. 2) the maximum level of
NOx conversion (associated with a C1:NOx ratio of ~4)
increased from ~70% (Oct. 7-8) to approximately 95% for the
runs on Oct. 16 and 17. The higher SV condition presented in
Fig 3 shows that the optimum C1:NOx ratio increased to ~7.5
and the conversion efficiency increased from 74% (on Oct. 8) to
around 85% for the Oct. 16 runs. This behavior is a strong
indication that the catalysts had not been fully aged prior to
evaluation. But, closeness of the data on Oct 16 and 17
suggests that catalyst aging was nearing completion on Oct. 16.
The cumulative engine running time put on the catalyst prior to
the Oct 16 runs was estimated to be around 32 hours. The
degreening period that was used was based on input from
catalyst chemists experienced with these catalyst systems.
However, the exceptionally low fuel sulfur level of the ECD-1
fuel (<15 ppm) likely necessitated a longer aging period to
ensure complete sulfate accumulation on the silver catalyst
sites.
SV ~ 21,000/h
2.5
2.0
SV ~ 57,000/h
1.5
1.0
0.5
60
50
0.0
Oct. 7-8: ECD1 fuel, Reductant = FG ethanol (untreated)*
Oct. 8-9: ECD1 fuel, Reductant = FG ethanol (untreated)
Oct. 10: ECD1 fuel, Reductant = FG ethanol (untreated)
Oct. 10: ECD1 fuel, Reductant = FG ethanol (distilled)
Oct. 15: ECD1 fuel, Reductant = FG ethanol (distilled)
Oct. 16: E-diesel fuel, Reductant = FG ethanol (distilled)
Oct. 17: ECD1 fuel, Reductant = FG ethanol (untreated)
Oct. 17: ECD1 fuel, Reductant = reagent grade ethanol
40
30
20
10
0
0
2
4
6
8
10
0.0
2.0
4.0
6.0
C1/NOx Ratio
8.0
10.0
12.0
Fig 4. Fuel penalty versus C1:NOx ratio for the two main engine
conditions.
HYDROCARBON SLIP MEASUREMENTS
12
C1(2X ethanol):NOx Ratio
Fig 2. NOx conversion as a function of C1:NOx ratio for the low
speed engine operating condition (1115 RPM, 230 ft-lb,
~21000/h space velocity. *Note, for this condition SV
10500/h.
As shown in Table 2 the engine-out hydrocarbon emissions
were low (< 50 ppm) for both operating conditions though there
was a slight increase associated with using E-diesel as the fuel.
These levels decreased further through the catalyst (when no
reductant was added), thus indicating that some oxidation was
taking place. However, the hydrocarbon emissions downstream
of the catalyst increased with the addition of ethanol into the
exhaust as depicted in Figs 5 and 6. (Note that the test
conditions presented in Figs 5 and 6 correspond to those in Figs
2 and 3 respectively.)
5
800
800
As with the data presented in Figs 2 and 5, the increase in slip
associated with succession of experimentation prevents detailed
assessment of HC slip. However, on Oct. 17 the HC slip was
exceptionally low for the runs utilizing high purity reagent-grade
ethanol as the reductant. This suggests that the impurities
present in the fuel-grade ethanol may contribute substantially to
HC slip (even if they do not influence NOx conversion).
600
600
TEMPERATURE EVALUATION
400
400
200
200
1200
1200
1000
C1(2X ethanol):NOx ~6.3
1000
HC Slip (ppm)
C1(2X ethanol):NOx ~2.8
0
200
0
400
350
300
250
The conversion efficiency of this particular ethanol-SCR system
is thought to be near optimum for catalyst temperatures near
400oC. However, during actual engine operation, the exhaust
temperatures can be much lower, particularly during idling and
low-load operation. Therefore we felt some measure of the low
temperature performance of ethanol-SCR was needed. The
results of this evaluation are shown in Fig. 7.
Catalyst Core Temperature (oC)
100
Fig 5. HC slip as a function of C1:NOx ratio for SV ~ 21000/h.
C1(2X ethanol):NOx ~6.3
250
Oct. 7-8: ECD1 fuel, Reductant = FG ethanol (untreated)*
Oct. 8-9: ECD1 fuel, Reductant = FG ethanol (untreated)
Oct. 10: ECD1 fuel, Reductant = FG ethanol (untreated)
Oct. 10: ECD1 fuel, Reductant = FG ethanol (distilled)
Oct. 15: ECD1 fuel, Reductant = FG ethanol (distilled)
Oct. 16: E-diesel fuel, Reductant = FG ethanol (distilled)
Oct. 17: ECD1 fuel, Reductant = FG ethanol (untreated)
Oct. 17: ECD1 fuel, Reductant = reagent grade ethanol
200
HC Slip (ppm)
NOx Conversion (%)
80
150
C1(2X ethanol):NOx ~2.8
60
40
20
0
200
100
250
300
350
Catalyst Core Temperature (oC)
400
Fig 7. NOx conversion as a function of temperature.
50
0
0
2
4
6
8
10
12
C1(2X ethanol):NOx Ratio
Fig 6. HC slip as a function of C1:NOx ratio for SV ~57000/h.
These figures also show that HC slip increased with succeeding
experiments up to the runs on Oct 16. This further supports the
assessment that the catalysts were insufficiently aged at the
end of the de-greening period. These catalyst systems are
believed to behave more like an oxidation catalyst in the green
state; i.e. they oxidize HC species present in the exhaust.
However, recent studies have indicated that as the silver sites
become sulfated, these catalysts become more selective to NOx
conversion, and hence their ability to oxidize hydrocarbons is
reduced as the sulfate level increases (9).
During each test condition the HC emissions increase in an
approximate linear manner with the flow rate of ethanol. As
expected the slip was higher for the high-speed condition than
the low-speed condition since the residence time is much lower.
Assuming straight-line behavior, then for a C1:NOx ratio of 6,
the HC slip is ~170 ppm for the high-speed condition versus
~150 ppm for the low-speed condition.
Throughout this test, the engine was run at a fixed speed of
1115 RPM. The load was initially set to 230 ft-lbs to achieve a
catalyst temperature between 350 and 400oC. Next, the load
was gradually reduced to achieve temperatures approximating
350, 300, and 250oC. (Note that the SV dropped from 21000/h
to 17000/h as the catalyst core approached 250oC.) The
C1:NOx ratio was maintained near 6.3 and 2.8 for each
temperature setting. The NOx conversion associated with these
temperatures for both C1:NOx ratios is shown in Fig. 6.
The NOx conversion was observed to drop sharply with
decreasing catalyst temperature for the two C1:NOx ratios
studied. However, the two curves appear to converge near a
value of ~25% as the catalyst temperature approaches 250oC.
Without additional measurements it is unclear whether this
value represents a lower bound.
The hydrocarbon slip
corresponding with the conditions presented in Fig. 7 are shown
in Fig. 8. In this figure, the HC slip increases with decreasing
catalyst temperature for both conditions. However the rise is
much more dramatic for C1:NOx ratio of 6.3 versus 2.8. In fact
the HC slip for the low C1:NOx ratio only begins to rise sharply
when the catalyst core temperature is lowered to ~250oC. This
suggests that the additional ethanol slips past the catalyst.
6
C1:NOx ~6.3
300
1000
250
800
200
6
NOx @ 21000/h
NOx @ 57000/h
600
400
400
200
200
350
300
4
150
3
100
250
2
50
1
0
0
200
0
0
0
2
4
6
8
C1(2X ethanol):NOx Ratio
Catalyst Core Temperature (oC)
Fig 8. Hydrocarbon slip as a function of temperature.
5
N2O @ 57000/h
10
12
Fig 10. N2O emissions as a function of space velocity and
C1:NOx ratio.
FTIR RESULTS AND BAG ANALYSIS
FTIR Results
A Nicolet Magna-IR 560 FTIR spectrometer equipped with a
REGA 700S sampling system was used to speciate NOx, N2O,
ammonia, and acetaldehyde emissions from the catalyst exit as
a function of SV and C1:NOx ratio. The results for ammonia,
N2O, and acetaldehyde and their corresponding NOx emissions
are shown in Figs. 9, 10, and 11 respectively.
The N2O emissions were also observed to increase with
increasing C1:NOx ratio (see Fig. 10); the concentrations of
N2O, however, are quite low (<10 ppm). This represents an
important finding since early HC-SCR catalyst formulations
emitted large concentrations of N2O during NOx conversion.
Like the ammonia results, the N2O emissions were higher for
the low-speed condition indicating that residence time is critical
for N2O formation. The kinetics of N2O formation and its role in
NOx reduction are not well understood.
300
250
NOx Level (ppm)
70
NOx @ 21000/h
NOx @ 57000/h
NH3 @ 21000/h
NH3 @ 57000/h
60
50
200
40
150
30
100
20
50
10
0
0
0
2
4
6
8
10
NOx @ 21000/h
NOx @ 57000/h
Acet @ 21000/h
Acet @ 57000/h
200
50
40
150
30
100
20
50
10
0
Ammonia Level (ppm)
300
60
250
NOx Level (ppm)
As shown in Fig. 9, ammonia was detected by the FTIR
spectrometer and was observed to increase with increasing
C1:NOx ratio. Interestingly, the ammonia emissions were
significantly higher for the low-speed condition than for the highspeed condition. This is an important finding since ammonia
formation and its subsequent use to further aid in NOx reduction
is one of the pathways that is believed to occur for NOx
reduction by silver-loaded alumina catalysts (3). The lower
residence time allows more ammonia to be formed, which
subsequently may be important for the conversion of NOx to N2.
12
C1(2X ethanol):NOx Ratio
Fig 9. Ammonia emissions as a function of space velocity and
C1:NOx ratio.
Acetaldehyde Level (ppm)
HC Slip (ppm)
600
400
7
N2O @ 21000/h
C1:NOx ~2.8
800
NOx Level (ppm)
1000
1200
N2O Level (ppm)
1200
0
0
2
4
6
8
10
12
C1(2X ethanol):NOx Ratio
Fig 11. Acetaldehyde emissions as a function of space velocity
and C1:NOx ratio.
The emissions of acetaldehyde (shown in Fig. 11) were
observed to increase with increasing C1:NOx ratio and reached
appreciable levels for the high-speed setting at moderate
C1:NOx ratios.
Acetaldehyde emissions are formed directly
from the oxidation of ethanol (via catalysis) and therefore would
be expected to increase proportionally with ethanol
concentration.
The much higher acetaldehyde emissions
associated with the high-speed condition suggests that 1) most
if not all the ethanol is converted into acetaldehyde over the
catalyst, and 2) the subsequent reactions utilizing acetaldehyde
to reduce NOx proceed less rapidly than does acetaldehyde
formation.
These results agree with the bench studies
performed by Noto et al. (5) which showed that most of the
ethanol is converted to acetaldehyde for these catalyst types.
7
Bag Analysis
The bag samples were collected using the miro-diluter
highlighted in the Fig. 1 schematic. CG-MS analysis revealed
that detectable levels of nitrophenol were formed during the
ethanol-SCR process. Although we were not able to accurately
quantify nitrophenol emissions, we are able to note that its
concentration increased with ethanol delivery and space
velocity.
CONCLUSIONS
A full-scale ethanol-SCR system demonstrated excellent
reduction of NOx emissions from a heavy-duty diesel engine for
exhaust and catalyst temperatures of 350-400oC. The NOx
conversion reached 95% to 85% for space velocities of 21000/h
and 57000/h respectively. In addition, the C1:NOx ratios used
to achieve these efficiencies were considered to be reasonable;
approaching a value of 4 for 21000/h condition and around 7 for
the 57000 condition.
This represents energy based fuel
penalties of approximately 1.5 and 1.8 % respectively. The NOx
conversion efficiency, however, depended greatly upon the
catalyst core temperature. When the core temperature was
lowered from ~360oC to ~250oC, the conversion efficiencies fell
to near 25% for both speed conditions.
Because the catalysts were not thoroughly aged prior to testing,
the NOx conversion improved with progression of
experimentation and thus interfered with data interpretation.
The results suggest that the de-greening period needed to be
extended to over 35 hours when using ECD-1 (and probably
any low sulfur fuel) as the fuel type. The increase in NOx
conversion with succeeding experimentation was accompanied
by an increase in HC slip which further supports the belief that
the accumulation of sulfate on the silver catalyst sites actually
enhances the NOx conversion for these particular SCR systems.
This investigation also showed that the ethanol was rapidly
converted to acetaldehyde by the silver-loaded alumina catalyst,
which subsequently slipped past the catalyst at appreciable
levels when the space velocity was increased to 57000/h.
Ammonia and N2O were also detected in the exhaust. The
concentrations for both these compounds were observed to be
higher for the low space velocity condition. However, the
concentration of N2O did not exceed 10 ppm, which is
considered to be exceptionally low.
ACKNOWLEDGMENTS
This work was supported by the Department of Energy
Office of Freedom CAR & Vehicle Technologies. Cummins
Engine Company, Inc. provided both the test engine and
auxiliary components.
The authors would also like to
acknowledge Dave Loos and Norm Marek of the Illinois
Department of Commerce and Community Affairs for their
efforts to coordinate the delivery of the splash-blended E-diesel
and other fuel components. Finally we would like to thank
Glenn Kenreck of GE Betz for performing a blending agent
compatibility study and Jeff Chambers (of ORNL) for assistance
with engine setup.
REFERENCES
1. Adelman, B., Literature Survey to Assess the State-ofthe-Art of Selective Catalytic Reduction of Vehicle NOx
Emissions, CRC Project No. AVFL-7, Archive C02-2814, June
21, 2002.
2. Majewski, W. A., Selective Catalytic Reduction,
DieselNet Technology Guide, Copyright Ecopoint Inc. Revision
2002.05a.
3. Burch, R., Breen, J. P., Meunier, F. C., A Review of the
Selective Reduction of NOx with Hydrocarbons Under LeanBurn Conditions with Non-Zeolite Oxide and Platinum Group
Metal Catalysts, Applied Catalysis B: Environmental 39 (2002)
382-303.
4. Eranen, K., Lindfors., Niemi, A., Elfving, P., and Cider,
L., Influence of Hydrocarbons on the Selective Catalytic
Reduction of NOx over Ag/Al2O3 – Laboratory and Engine
Tests, SAE Paper No. 2000-01-2813.
5. Noto, T., Murayama, T., Tosaka, S., and Fujiwara, Y.,
Mechanism of NOx Reduction by Ethanol on a Silver-Base
Catalyst, SAE Paper No. 2001-01-1935.
6. Satokawa, S., Yamaseki, K., and Uchida, H., Influence
of a Low Concentration of SO2 for Selective Reduction of NO by
C3H8 in Lean-Exhaust Conditions on the Activity of Ag/Al2O3
Catalyst, Applied Catalysis B: Environmental 34 (2001) 299306.
7. Thomas, S. E., Shawcross, J. T., Gillespie, R.,
Raybone, D., and Martin, A. R., The Role of NO Selective
Catalysts in the Plasma Enhanced Removal of NOx and PM
from Diesel Exhausts, SAE Paper No. 2001-01-3568.
8. Christopher L. Aardahl (Pacific Northwest National
Laboratory. Personal communication
9. Paul
W.
Park
(Caterpillar
Inc.).
Personal
communication.
10. Park, P. W., Boyer, C. L., Ragle, C. S., Balmer, M. L.,
Aardahl, C. L, Rappe, K. G., and Tran, D. N., Lean-NOx
Catalyst Development for Diesel Engine Applications, Presented
at the 2002 DEERE Conference, SanDiego, CA, Aug 25-29,
2002.
11. Satokawa, S., Yamaseki, K., and Uchida, H., Influence
of a Low Concentration of SO2 for Selective Reduction of NO by
C3H8 in Lean-Exhaust Conditions on the Activity of Ag/Al2O3
Catalyst, Applied Catalysis B: Environmental 34 (2001) 299306.
12. W. H. Lane, The Caterpillar DeNOx Catalyst System,
Company Brochure.
13. Kass, M. D., Thomas, J. F., Storey, J. M., Domingo, N.,
Wade, J., and Kenreck, G.; Emissions From a 5.9 Liter Diesel
Engine Fueled with Ethanol Diesel Blends, SAE Paper No.
2001-01-2018, 2001.
8